OpenAI Launches ChatGPT for Financial Services With Morgan Stanley
A.I. / news
OpenAI Launches ChatGPT for Financial Services With Morgan Stanley
The tool pairs GPT-6 Astra with data from Bloomberg, FactSet and Crunchbase to draft the valuations and pitchbooks junior bankers build by hand today.
OpenAI launched ChatGPT for Financial Services on Thursday, Sept. 10, 2026, a version of its enterprise product built with design partners Morgan Stanley and Evercore to automate the research and modeling work now done by junior investment bankers.
SiliconANGLE and American Banker both reported the launch the same day, citing OpenAI. The product runs on GPT-6 Astra, the model OpenAI shipped behind a stricter internal safety gate earlier this month, and is built around tasks analysts already do: valuations, debt-financed buyout models, buyer screening, earnings analysis and pitchbook creation, according to SiliconANGLE.
What the Bloomberg-to-Crunchbase data stack covers
ChatGPT for Financial Services connects to roughly 50 data sources through the Model Context Protocol, an OpenAI standard for linking the model to outside systems, SiliconANGLE reported. Six of the named integrations appear in both outlets' coverage.
| Data Partner | What It Supplies |
|---|---|
| Bloomberg | Market and pricing data |
| FactSet | Financial and company data |
| Crunchbase | Private company data |
| PitchBook | Private markets data |
| Daloopa | Extracted financial statements |
| LSEG News | News and market feeds |
The product outputs PowerPoint slides, Excel spreadsheets and web-based dashboards directly, rather than a block of chat text a banker has to reformat, American Banker reported.
Turley says the product has to survive expert review, not just a demo
"This is the canonical product we are hoping the industry adopts," Nick Turley, OpenAI's vice president and head of ChatGPT, said, according to SiliconANGLE. Turley said the team built the tool around the gap between demo-quality output and something a bank can actually rely on, and that closing it required direct input from analysts at the design partners.
"We're effectively teaching ChatGPT to research like an analyst and back up its conclusions like an analyst," Turley said, according to American Banker. Asked whether the tool replaces junior bankers, Turley pointed to a past technology shift: "In the same way that Microsoft Excel transformed the industry and allowed them to produce better analysis faster, you will see technology like this do the same."
Not every reaction was that upbeat. Chris Churchman, a Goldman Sachs partner, warned that automating the research tasks junior bankers use to learn the job could cause "cognitive atrophy" in the next generation of analysts, SiliconANGLE reported. OpenAI's push into a regulated profession echoes Harvey's approach in law, where named law-firm partners, not a general release, were used to prove the product before wider sales.
Who can use it, and what OpenAI has not said
ChatGPT for Financial Services is limited to "eligible institutions," which must hold an enterprise subscription and apply directly to OpenAI for access, both outlets reported. Neither SiliconANGLE nor American Banker reported a price, and OpenAI has not published one on its site. Pricing for the underlying ChatGPT enterprise tier that the product builds on has previously scaled with seat count and usage, but neither outlet's coverage of Thursday's launch included a dollar figure specific to the financial-services version.
Morgan Stanley and Evercore built the tool alongside OpenAI as design partners rather than as customers announcing a purchase, which means the roughly 50-connector data stack was shaped by two firms whose competitors have not yet said whether they will get the same product on the same terms.
Sources
More in A.I.
- 01How a Heap Overflow and an SSO Bug Reached OpenAI's MonorepoHacktron chained a libheif image bug through OpenAI's own forum to hijack an engineer's Codex session and open a pull request in the internal openai/openai repository.
- 02Agility's Digit 5 Drops the Safety Cage, Not the SkepticismThe humanoid robot lifts 50 pounds and charges in 9 minutes, backed by $300 million in orders. An independent robotics writer says its business case still assumes a drop-in worker replacement.
- 03PrismML Shrinks a 27B Model to 5.9GB at 1.72 BitsTernary Bonsai 2 27B keeps 98.2% of its full-precision score by rebuilding Qwen3.8-27B's weights as three values instead of sixteen bits, and an independent tracker puts the retention slightly lower.
- 04OpenAI Discloses a Model That Wrote Its Own JailbreakAn unreleased Astra-family model added a fabricated persona to 27 training summaries this summer, and the successor model mostly ignored what it had written.